Emerging data suggest that maternal vitamin A deficiency may be a risk factor for mother to child transmission of HIV. Among HIV infected women, low serum etinol concentration during pregnancy is a risk factor for infant mortality, mother to child transmission, and higher concentrations of breast milk HIV DNA detected by PCR compared to HIV infected women with normal serum retinol concentrations. These observations suggest that vitamin A supplementation of HIV infected women may reduce vertical HIV transmission rates. Several large trials are currently under way or planned to test this hypothesis. The interaction of vitamin A with HIV, however, appears complex. Some in-vitro data indicates that vitamin A supplementation may benefit the HIV itself. In several studies, pre-treatment of human monocytes or monocyte-like cells with retinoic acid (the intracellular form of vitamin A) prior to infection with HIV increased virus production, and in one study retinol (the form in circulations) and retinyl acetate (the form in food and supplements) produced similar effects leading these authors to conclude that "the use of retinoids in HIV infected patients should be used with caution." We conducted this randomized placebo controlled trial among 40 HIV seropositive women to carefully monitor viral load and other potentially adverse immune responses over close intervals following a single oral dose of 300,000 IU vitamin A or placebo. Baseline studies carried out on the day before allocation to treatment included viral load, serum retinol, total lymphocytes and selected lymphocyte subsets, and activation markers. On the day following treatment and at weeks 1, 2, 3, 4, and 8 immune response measures were repeated. At Day 1 and Week 1 women were queried for possible side effects. Results demonstrated that the vitamin A dose was well tolerated with no side effects. There were no significant differences in treatment group means or medians of any lymphocyte subset measured, in the CD4/CD8 ratio or in total lymphocyte count at any time point. Geometric mean viral load and median viral load was not different between treatment groups at any time point. There did appear to be greater variability in viral load in both directions among vitamin A compared to placebo recipients. Women who experienced one or more increases in viral load of 0.5 log 10 or more following vitamin A were distinguished from those who did not as having significantly lower baseline viral load concentrations. There was no distinct pattern of viral load changes among the vitamin A women, and an equal number experienced increase and decreases following vitamin A. These findings do not provide any evidence of an adverse effect of high dose vitamin A supplementation and provide reassurance that potential beneficial effects of vitamin A in HIV infection can be studied with little apparent risk.